Medallion economics

Sam Morgan tweets (1, 2):

Sydney Taxi Economics: 2 drivers, 24hrs a day. Owner of $170,000 taxi license rents car for $1550 per week. Revenue ~$3500 per week. $1000 income per driver per week.

$1000/((24*7)/2) = 84 hours. $11.90 per hour. License owner yields $80,600 p.a. on $170,000 licence.

I'm not sure from where Sam sourced his figures, but if they're right, that's a very good annual return on investment. Sufficiently high that I wonder why more folks don't invest in medallions.

Let's work some of that back though. Suppose a car costs $60k and fully depreciates after three years if driven 24/7. So deduct $20k per annum from the revenue stream. Maybe another $7.5k in maintenance, vehicle registration fees, and car insurance [all of these are just guesses]. But even that still keeps us north of $50k p.a. on $170k licence. Add in a bit in management costs for the license-holder: ensuring the drivers aren't beating up the vehicle too much, replacing drivers if one leaves and so on. Is the rest regulatory uncertainty or have I missed some costs facing license-holders? Or, to put it another way, should I be trying to buy taxi licenses?